(c)2023 UW Math Dept licensed under a Creative Commons By-NC-SA 4.0 International License. Consequences of the rules of probability 1. (Warm-up) We sample a point (𝑥, 𝑦) uniformly from [0, 1] × [0, 2]. Let Ω = [0, 1] × [0, 2]. Compute the probability of the following events: (a) 𝐴 = {(𝑥, 𝑦) ∈ Ω : 𝑥 > 𝑦} Solution note: 1 4 (b) 𝐵 = {(𝑥, 𝑦) ∈ Ω : 𝑥 ≤ 𝑦} Solution note: 3 4 (c) 𝐶 = {(𝑥, 𝑦) ∈ Ω : 𝑥 < 21 } Solution note: 1 2 (d) 𝐷 = {(𝑥, 𝑦) ∈ Ω : 𝑥 > 34 } Solution note: 1 4 2. (Quick check) Consider the same set-up as Problem 1. Use set operations to express each event below in terms of the events from Problem 1. There may be multiple answers. (a) The event Ω. Solution note: We have Ω = 𝐴 ∪ 𝐵. Importantly here: 𝐴 and 𝐵 are complements of each other, and in general, 𝐴 ∪ 𝐴𝑐 = Ω. (b) The event that both 𝑥 > 𝑦 and 𝑥 > 43 . Solution note: This event can be expressed as 𝐴 ∩ 𝐷. (c) The event that 1 2 ≤ 𝑥 ≤ 34 . Solution note: This event can be expressed as 𝐶 𝑐 ∩ 𝐷 𝑐 or (𝐶 ∪ 𝐷)𝑐 . From (a), we also have Ω = 𝐴 ∩ 𝐵 so that we also have that this event is (𝐴 ∩ 𝐵) \ (𝐶 ∪ 𝐷). 3. We flip a fair coin until we have seen at least one heads and one tails and then we stop. We had previously modeled this experiment with sample space Ω = {2, 3, . . . , } ∪ {∞} and with probabilities P({∞}) = 0 and P({𝑘 }) = 1 , 2𝑘−1 𝑘 = 2, 3, . . . . We want to compute the probability of flipping the coin at least 4 times. Letting 𝐴 be this event of interest, we do this in multiple steps: (a) Describe the event 𝐴𝑐 and express P(𝐴) in terms of P(𝐴𝑐 ). Solution note: The event 𝐴𝑐 is the event that we flip the coin only 2 or 3 times. Since Ω = 𝐴 ∪ 𝐴𝑐 with 𝐴 and 𝐴𝑐 disjoint, we have 1 = P(Ω) = P(𝐴) + P(𝐴𝑐 ) =⇒ P(𝐴𝑐 ) = 1 − P(𝐴), by countable additivity. (b) Now decompose 𝐴𝑐 into two simpler events. Compute the probability of the two simpler events. Solution note: We have 𝐴𝑐 = {2, 3} = {2} ∪ {3} with P({2}) = 1 2 and P({3}) = 41 . (c) Use your answer in (b) to compute P(𝐴𝑐 ) Solution note: We again using countable additivity: P(𝐴𝑐 ) = P({2, 3}) = P({2}) + P({3}) = 1 1 3 + = . 2 4 4 (d) Put your answer in (a) and (c) together to compute P(𝐴). Solution note: Finally, we arrive at P(𝐴) = 1 − P(𝐴𝑐 ) = 1 − 3 1 = . 4 4 (e) How do you think the probability of flipping the coin at least 6 times compares to your answer in (d)? Explain. Solution note: It will be smaller. The event 𝐴 can be decomposed into the union of two disjoint events: the event 𝐵 where there are at least 6 coin flips and the event 𝐶 where there are 4 or 5 coin flips. Since P(𝐶) must be ≥ 0, then P(𝐴) = P(𝐵 ∪ 𝐶) = P(𝐵) + P(𝐶) ≥ P(𝐵) by countable additivity. 4. In our version of Bingo, balls numbered 1 through 25 are removed, one at a time, from a rotating cage. After 2 balls have been removed, we want to know our chances that of seeing a square (e.g. 4) or a cube (e.g., 8). We do this in several steps. (a) What is the chance of not seeing a square? What is the chance of seeing a square? Solution note: If we do not keep track of the order in which balls are selected, then there are 25×24 2 possible outcomes. Among these outcomes, there are (25 − 5)(25 − 6) 2 outcomes in which we don’t see a square. (Note: there are 5 squares between 1 and 25.) Thus, the chance of not seeing a square is: 20 × 19 . 25 × 24 Using this answer, the chance of seeing a square is thus: 1− 20 × 19 . 25 × 24 (b) What is the chance of not seeing a cube? What is the chance of seeing a cube? Solution note: Using same logic from (a) and given there are 2 cubes between 1 and 25, the chance of not seeing a cube is (25 − 2) × (25 − 3) 23 × 22 = . 25 × 24 25 × 24 The chance of not seeing a cube is then 1− 23 × 22 . 25 × 24 (c) What is the chance of seeing a square and a cube? Solution note: There are 5 squares and 2 cubes. The number 1 is both a cube and a square. Therefore, we have 24 outcomes in which 1 is selected, and, we have 4 outcomes where a square and the cube other than 1 is selected. This yields 28 outcomes, and the chance of seeing a square and a cube is 28 56 = 25 × 24/2 25 × 24 (d) Use your answers from (a)–(c) to compute the probability of seeing a square or a cube. Hint. Use the inclusion-exclusion formula: P(𝐴 ∪ 𝐵) = P(𝐴) + P(𝐵) − P(𝐴 ∩ 𝐵) Solution note: Let 𝐴 be the event of seeing a square and 𝐵 be the event of seeing a cube. In (c), we considered event 𝐴 ∩ 𝐵. We are interested in 𝐴 ∪ 𝐵. Using the hint, we have P(𝐴 ∪ 𝐵) = P(𝐴) + P(𝐵) − P(𝐴 ∪ 𝐵) 20 × 19 23 × 22 56 =1− +1− − 25 × 24 25 × 24 25 × 24 20 × 19 + 23 × 22 + 56 =2− 75 × 74 Consequences. To summarize, we have several strategies for using use set operations to compute the probability of events. These include: • Complements: P(𝐴) = 1 − P(𝐴𝑐 ) Í • Decomposition: P(𝐴) = 𝑖 P(𝐴𝑖 ) for countable disjoint 𝐴𝑖 with 𝐴 = ∪𝑖 𝐴𝑖 . • Monotonicity: 𝐴 ⊆ 𝐵 implies P(𝐴) ≤ P(𝐵) • Inclusion-exclusion for 2 events: P(𝐴 ∪ 𝐵) = P(𝐴) + P(𝐵) − P(𝐴 ∩ 𝐵) or 3 events: P(𝐴 ∪ 𝐵 ∪ 𝐶) = P(𝐴) + P(𝐵) + P(𝐶) − P(𝐴 ∩ 𝐵) − P(𝐵 ∩ 𝐶) − P(𝐴 ∩ 𝐶) + P(𝐴 ∩ 𝐵 ∩ 𝐶) or n events: P(∪𝑛𝑖=1𝐴𝑖 ) = 𝑛 ∑︁ 𝑘=1 (−1)𝑘+1 ∑︁ 𝑖 1 <𝑖 2 <...<𝑖𝑘 P(𝐴𝑖 1 ∩ 𝐴𝑖 2 ∩ . . . ∩ 𝐴𝑖𝑘 ).